Anda di halaman 1dari 10

Received: 20 April 2018 Revised: 28 July 2018 Accepted: 4 September 2018

DOI: 10.1111/pedi.12777

ORIGINAL ARTICLE

Psychosocial benefits of insulin pump therapy in children with


diabetes type 1 and their families: The pumpkin multicenter
randomized controlled trial
Esther Mueller-Godeffroy1 | Reinhard Vonthein2,3 | Carmen Ludwig-Seibold4 |
Bettina Heidtmann5 | Claudia Boettcher6 | Miriam Kramer6 | Nicole Hessler2 |
Doerte Hilgard7,8 | Eggert Lilienthal9 | Andreas Ziegler2,3,10 | Verena M. Wagner1,11 |
on behalf of the German Working Group for Pediatric Pump Therapy (agip)

1
Department of Pediatric and Adolescent
Medicine, University of Luebeck, Luebeck, Objective: Continuous subcutaneous insulin infusion (CSII) is on the rise among pediatric
Germany patients with type 1 diabetes mellitus. Metabolic effects alone cannot explain this rising popu-
2
Institute of Medical Biometry and Statistics, larity. From the patient's perspective, the main benefits of CSII may be found in subjective psy-
University of Luebeck, University Medical
chosocial health outcomes (patient-reported outcomes [PRO]).
Center Schleswig-Holstein, Luebeck, Germany
3
Subjects and Methods: In a multicenter open randomized controlled trial, children and adolescents
Center for Clinical Trials, University of
Luebeck, Luebeck, Germany aged 6 to16 years currently treated with multiple daily injections (MDI) were randomized 1:1, stratified
4
Department of Pediatrics, by center, to either starting with CSII immediately after the baseline interview or to continuing MDI
Oberschwabenklinik GmbH, Wangen, while waiting 6 months for transmission to CSII. The primary outcomes were patient-reported diabetes-
Germany specific health-related quality of life (DHRQOL) and diabetes burden of the main caregiver. Secondary
5
Catholic Children's Hospital, Wilhelmstift outcomes were caregiver stress, fear of hypoglycemia, satisfaction with treatment, and HbA1c.
gGmbH, Department of Endocrinology and
Diabetology, Hamburg, Germany
Results: Two-hundred and eleven patients were randomized between February 2011 and
6 October 2014, and 186 caregivers and 170 patients were analyzed using the intention-to-treat
Division of Pediatric Endocrinology and
Diabeteology, Center of Child and Adolescent principle for primary outcomes. Children 8 to 11 years in the CSII group reported improved
Medicine, Justus Liebig University Giessen, DHRQOL at follow-up compared to MDI (median difference [MD] 9.5, 95% confidence interval
Frankfurt, Germany
[CI] 3.6-16.7, P = 0.004). There were no treatment differences in the adolescent age-group
7
Department of Pediatrics,
12 to 16 years (MD 2.7; 95% CI −3.2-9.5; P = 0.353). The main caregivers of the CSII group
Gemeinschaftskrankenhaus Herdecke gGmbH,
Herdecke, Germany reported a significant decline of overall diabetes burden at follow-up compared to the MDI
8
Pediatric and Adolescent Medical Practice, group (MD 0; 95% CI −1-0; P = 0.029). Secondary PROs also were in favor of CSII.
Witten, Germany Conclusions: CSII has substantial psychosocial benefits. PROs demonstrate these benefits.
9
Department of Pediatrics, St. Josef-Hospital, Registered as NCT01338922 at clinicaltrials.gov
Ruhr-University Bochum, Bochum, Germany
10
School of Mathematics, Statistics and KEYWORDS
Computer Science, University of KwaZulu-
Natal, Pietermaritzburg, South Africa children/adolescents, continuous subcutaneous insulin infusion (CSII), diabetes type 1, health-
related quality of life, randomized controlled trial

An abstract of less than 300 words was presented at the IPSAD 2017, Innsbruck, Austria.
Abbreviations: CSII, continuous subcutaneous insulin infusion; DCCT, diabetes control and complications trial; DDG, : german diabetes society (Deutsche Diabetes
Gesellschaft); DFCS, diabetes family conflict scale; DHRQOL, diabetes-specific HRQOL; DSMB, data safety and monitoring board; DTSQs, diabetes treatment satis-
faction questionnaire, status version; DTSQs-P, parent version of the DTSQ-s; DTSQs-T, teen version of the DTSQ-s; DTSQ-P TS+, summary score of the parent ver-
sion of the DTSQ-s; DTSQ-T TS+, summary score of the teen version of the DTSQ-s; FAS, full analysis set; HFS-P, hypoglycemia fear survey, parent version; HRQOL,
health-related quality of life (HRQOL); IMBS, Institute of Medical Biometry and Statistics, University of Luebeck; ISPAD, International Society for Pediatric and Ado-
lescent Diabetes; KINDL-DM, diabetes-specific module of the KINDL-R; KINDL-R, modular HRQOL questionnaire for children-–revised version; MDI, multiple daily
injections; PIP, pediatric inventory for parents; PIP-D, total difficulty score of the pediatric inventory for parents PIP; PIP-F, total frequency score of the pediatric
inventory for parentsPIP; PRO, patient- -reported outcomes; RCT, randomized-controlled trial; RITA, randomization in treatment arms; SAS software, statistical analy-
sis system software; SES, socioeconomic status; WHO-5, World Health Organization-five

© 2018 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd
Pediatric Diabetes. 2018;1–10. wileyonlinelibrary.com/journal/pedi 1
2 MUELLER-GODEFFROY ET AL.

11
Pediatric and Adolescent Medical Practice,
Rostock, Germany
Correspondence
Esther Mueller-Godeffroy, Department of
Pediatric and Adolescent Medicine, University
of Luebeck Ratzeburger Allee 160, 23538
Luebeck, Germany.
Email: esther.mueller-godeffroy@uksh.de
Funding information
Deutsche Forschungsgemeinschaft, Grant/
Award Number: WA 2929/ 2-1; Roche
Diabetes Care Germany GmbH, Grant/Award
Number: 501 000 9391

1 | I N T RO D UC T I O N an open RCT. We investigated the impact of CSII on the DHRQOL of


the patients and on the overall diabetes burden of the main caregivers
The ultimate goal in treating children and adolescents with type 1 dia- (primary outcomes). We hypothesized that children and adolescents
betes mellitus (T1DM) is to achieve optimal glycemic control, avoiding treated with CSII showed better DHRQOL compared to those treated
acute, and long-term complications, without compromising age- with MDI, and their caregivers less overall diabetes burden. We also
1
appropriate development and health-related quality of life (HRQOL). investigated the impact of CSII on generic HRQOL, caregiver stress,
Continuous subcutaneous insulin infusion (CSII) allows for a more fear of hypoglycemia, family conflict, treatment satisfaction, and
physiologic insulin replacement and is on the rise among pediatric HbA1c (secondary outcomes). We hypothesized better generic
patients.2 CSII may be superior to multiple daily injections (MDI) for HRQOL, less caregiver stress, less fear of hypoglycemia, less family
3,4
metabolic control and for acute complications. However, metabolic conflicts, a higher level of patient satisfaction, and a slightly lower
effects alone cannot explain the rising popularity of CSII.3,5 level of HbA1c in the CSII group compared to the MDI group.
To understand the true impact of a treatment regimen,
researchers should look beyond metabolic efficacy and examine the
families' experiences of their day-to-day living with diabetes.6 As
2 | METHODS
these are inherently subjective, they should be reported by the young
patients or caregivers themselves. Patient-reported outcomes (PRO), 2.1 | Subjects and study design
specifically HRQOL, are increasingly incorporated in clinical diabetes We performed an open multicenter parallel randomized controlled
research.7 intervention trial with waiting-list control group in children with
CSII provides more flexibility in lifestyle, which may positively T1DM and their main caregiver, usually a parent. Eighteen German
affect the HRQOL and disease burden. It is not yet established specialized pediatric diabetes centers approved by the German Diabe-
whether the increased flexibility translates into improved standardized tes Society (Deutsche Diabetes Gesellschaft, DDG) enrolled the
measures of PRO.8 Measures of PRO can be either generic or patients and conducted the interventions. All children and adolescents
diabetes-specific. Generic measures cover a broad range of subjective aged 6 to 16 years with T1DM currently being treated with MDI with
health aspects relevant for populations with and without health prob- an indication for transfer to CSII were eligible for the present study.
lems. Diabetes-specific measures are designed to cover meaningful Exclusion criteria were diabetes duration of less than 6 months, remis-
aspects of the disease and its treatment, and hence, offer a greater sion (less than 0.5 U insulin/kg body weight), and insufficient German
depth of insight to the experiences of patients. They are also more literacy to complete questionnaires. The ethics committees of the par-
sensitive to treatment differences.9,10 ticipating centers approved the study. The investigations have been
Systematic reviews of pump therapy in children and adoles- conducted in accordance with the principles of the Declaration of Hel-
cents3,5,11,12 report mixed results on the HRQOL-benefits of CSII in sinki (2008). Patients were grouped as young children (6-7 years),
the pediatric age-group. All in all, there are improvements in diabetes- school-children (8-11 years), and adolescents (12-16 years). Informed
specific rather than generic aspects of HRQOL favoring CSII over consent was given by adolescents and both parents; children
MDI, however, only with weak evidence.5,13 There is a lack of meth- assented. The trial was registered in clinicaltrials.gov as
odologically sound studies, particularly of adequately powered ran- NCT01338922.
domized controlled trials (RCT). Patients eligible for transfer to CSII were randomized to either
A multicenter prospective observational pilot study preceding the immediate transition to CSII or MDI and additional individual optimi-
pumpkin trial found increased diabetes-specific HRQOL (DHRQOL) zation of the diabetes regimen—to control for educational effects—
and decreased diabetes burden after transfer to CSII and identified and transfer to CSII after second interviews 6 months later. Indication
psychometrically sound instruments.14 for CSII followed the clinical criteria of the German Clinical Practice
The objective of this study was to investigate the effects of CSII Guidelines, namely, recurrent severe hypoglycemia, wide fluctuations
on PRO in children and adolescents with T1DM and their families in in blood glucose levels, suboptimal diabetes control, pronounced
MUELLER-GODEFFROY ET AL. 3

dawn−/ dusk phenomenon, needle phobia, and also improvement of The main caregivers' subjective mental well-being was
the young patient's motivation, the wish for more flexibility in meal- assessed using the World Health Organization-Five Well-Being
times and daily routine, and patients' preference (not specified). Index (WHO-5).22 5-point Likert-scaled items are combined to
Patients were transferred to CSII by a multidisciplinary diabetes give a summary score. Higher scores indicate greater well-
15 being.
team.
Parenting stress was measured with the pediatric inventory for
parents (PIP). The main caregiver rated each of the 42 items using a
2.2 | Instruments
5-point Likert scale as to both the items frequency and the level of
The main caregiver provided sociodemographic data. The family's difficulty associated with it. Items are combined to give a total fre-
socioeconomic status (SES) was measured with the Winkler-Index, quency score (PIP-F), and a total difficulty score (PIP-D).23 Higher
comprising the three components: education/vocational qualification, scores indicate more parenting stress.
occupational status, and net household income. The SES index was The main caregiver completed the Hypoglycemia Fear Survey,
categorized as low, moderate, or high.16 parent version (HFS-P).24 The HFS-P comprises 25 items on two
PRO were measured using standardized questionnaires. Children scales: Behavior (10 items) and Worry (15 items). The scales are sepa-
and adolescents 8 to 16 years self-reported on their HRQOL, adoles- rately reported as recommended,25 higher scores indicating higher
cents 12 years and above additionally reported on family conflicts, levels of fear of hypoglycemia.
and satisfaction with their diabetes therapy. The main caregivers All original versions of the questionnaires have been validated
proxy-reported on the HRQOL of children younger than 8 years; (for more detailed information see Ref. 14). With the exception of the
those of adolescents reported on family conflicts. All other caregiver- DTSQ, all questionnaires have been used in the pilot study preceding
reported outcomes were measured across all age-groups. the pumpkin trial and displayed satisfactory reliability, and sensitivity
to treatment change.14
2.3 | Primary outcomes The multiple of the mean method (MOM) was used to
mathematically standardize HbA1c-values to the diabetes control and
Children's and adolescents' DHRQOL was measured using the
complications trial (DCCT) reference range: 20.7 to 42.6 mmol/mol
diabetes-specific module of the KINDL-R modular HRQOL question-
(4.05%-6.05%).26
naire.17 The CSII-adapted version of the diabetes-specific module
comprises 21 Likert-scaled items for patients 8 years and older which
are summarized to give a diabetes-specific KINDL-DM total score.
2.5 | Statistical analysis
Higher scores indicate better DHRQOL. Analyses used the full analysis set (FAS) following the intention-to-
The main caregivers reported on their overall diabetes burden treat principle. A patient was included in the FAS if randomized to CSII
using a one-dimensional 5-point intensity scale. Higher scores indicate or MDII, regardless of which intervention the patient actually received
a greater diabetes burden. and regardless of other protocol deviations. Patients were excluded
from this dataset only for the following reasons: violation of inclusion
or exclusion criteria, withdrawal of consent after randomization, or
2.4 | Secondary outcomes
complete lack of data. Missing data were assumed to be missing
Children's and adolescents' generic HRQOL were measured with the completely at random, so that no imputation was needed. The three
generic core measure of the KINDL-R.17 The instrument provides age- primary endpoints (school-child DHRQOL, adolescent DHRQOL, and
appropriate self-report versions (24 items) for school-aged children main caregiver overall diabetes burden at 6 months) were tested using
and adolescents, and a parent-proxy-report version for younger chil- a fallback procedure at global significance level 0.05.27 The two
dren. The Likert-scaled items are summarized in the generic KINDL R hypotheses on all cases and on age group 8 to 11 years were each
total core. Higher scores indicate better generic HRQOL. assigned an adjusted level of 0.025. The fallback would be to either
Adolescents and their main caregivers completed the adapted hypothesis in the first step. If both were significant, the full level
version of the diabetes family conflict scale (DFCS). Nineteen items would fall to the hypothesis on the effect on adolescents. A support-
are rated on a 3-point Likert-scaled and summarized to give a DFCS ing figure gives a detailed description of the fallback procedure.
total score, higher scores indicating a higher level of family Endpoints were analyzed using the exact two-sided U test for
conflict.18,19 treatment comparisons and signed-rank tests for changes from base-
Adolescents and their main caregivers completed the Diabetes line within groups and Hodges-Lehmann 95% confidence interval (pri-
Treatment Satisfaction Questionnaire, status version (DTSQs Parent, mary analysis). Rank-based analyses of covariance with baseline
14 items, and DTSQs Teen, 12 items), the German version being vali- adjustments for the respective dependent variable, age as a covariate
dated in the pumpkin study. Items on satisfaction with treatment and (when more than one age group was involved), and site as a random
metabolic control are summarized in the DTSQ-P TS + C (caregivers) effect yielded additional descriptive P values.
and the DTSQ-T TS + C (adolescents) summary scores. The original Expected complications (severe hypoglycemia and ketoacidosis)
instrument has shown good psychometric properties, validity and sen- and other adverse events were listed.
sitivity to change.20,21 Higher scores indicate greater satisfaction with All analyses were performed with SAS software (version 9.2, SAS
treatment. Institute, Cary, North Carolina, USA).
4 MUELLER-GODEFFROY ET AL.

2.6 | Sample size 3.2 | Primary and secondary outcomes


Sample size calculation was performed by using empirical proportions Table 2 presents the primary outcomes—self-reported DHRQOL of
of age-groups, and means and SD for the primary endpoints observed the school children and adolescents, and overall diabetes burden of
in the pilot study.14 The significance level (two-sided) for the testing the main caregiver—at baseline and at 6 months of CSII or MDI.
of the primary hypotheses was set to 5%, and each hypothesis was to As test 2 of DHRQOL in school children was significant
be tested at the 2.5% level. The power was set to 80%, and the drop- (P = 0.004) at level 0.025, and test 1 of parental burden was signifi-
out was assumed to be 20%. Sample size calculation for t-tests was cant (P = 0.029) at level 0.05, test 3 was conducted at level 0.05. This
corrected by 5% for U-tests. For the age distribution in the pilot study, test of treatment effect on DHRQOL in adolescents was not signifi-
2 × (32 + 26 + 78) = 272 patients of all age groups had to be initially cant (P = 0.353), ending the confirmatory procedure.
recruited. As age groups were more balanced at data safety and moni- School children (8-11 years) of the CSII group reported signifi-
toring board review, a total sample size of about 210 patients was cantly better DHRQOL at follow-up compared to the MDI group. The

considered sufficient. The recruitment was immediately amended. The difference between the groups was statistically significant in the pri-

power of the tests to detect differences in both subgroups given mary and the adjusted analysis. In adolescents (12-16 years), there

DHRQOL values from the pilot study was 96% in school-children and was no difference between the treatment groups detected. The analy-

77% in adolescents. The significance-level (two-sided) for the descrip- sis of the total sample (without age stratification) revealed a signifi-

tion of secondary outcomes was set to 1%. Neither adaptations nor cantly better DHRQOL at follow-up in favor of CSII (MD 5.95
[1.19-10.71], P = 0.016).
interim analyses were planned.
The main caregivers of the children and adolescents (6-16 years)
of the CSII group reported a significantly lower overall diabetes bur-
2.7 | Randomization den at follow-up compared to those of the MDI group. The difference
was statistically significant in both the primary and the adjusted
The randomization sequence was generated by validated software
analysis.
RITA (randomization in treatment arms) as stratified by center with a
Table 3 presents secondary patient- and caregiver-reported out-
1:1 allocation using permuted blocks with variable confidential block
comes, and HbA1c- values.
lengths two and four. Randomization results were provided by telefax
There were no significant changes in patient-reported generic
after registration of the patient at the Institute of Medical Biometry
HRQOL, and family conflicts. Adolescent-reported treatment satisfac-
and Statistics, University of Luebeck. This guaranteed concealment of
tion (DTSQ-T TS+) improved in both treatment groups. The difference
allocation.
in favor of CSII (P < 0.05) failed to reach the predefined significance
level of 1% for secondary outcomes in both the primary and the
adjusted analysis.
3 | RESULTS Improvement in subjective well-being of the main caregivers did
not differ significantly between the treatment groups. The main care-
3.1 | Recruitment and participants givers of the CSII group reported significantly reduced parental stress
(PIP-F, PIP-D), reduced fear of hypoglycemia (HFS Behavior, HFS
Between February 2011 and October 2014, 367 patients were
Worry), and improved treatment satisfaction (DTSQ-P, TS+C) at
assessed for eligibility in the clinical trial. In total, 211 patients were
follow-up compared to the MDI group. The differences between the
randomized, 106 to CSII and 105 to MDI.
treatment groups were significant in both the primary and the
Two and five patients, from CSII and MDI, respectively, did not
adjusted analysis with the exception of parenting stress, which failed
receive the intended intervention, 14 and 16 patients, respectively,
to reach the significance level of 1% in the unadjusted analysis
were lost to follow-up. (Figure 1) The FAS of 199 patients comprised
(P < 0.05). There were no differences between the treatment groups
data of 179 for primary endpoints.
regarding family conflict (DFCS).
Table 1 presents the demographic and clinical characteristics of
Children and adolescents displayed no significant change of
the study participants.
HbA1c in the study period independently of treatment regime. There
Mean age at randomization, the proportion of male patients,
was no significant decrease in severe hypoglycemia ISPAD II (6/3 and
mean diabetes duration, and the distribution of SES was similar in
2/2 at baseline and follow-up in CSII/ MDI group), ISPAD III (3/4 and
both groups. Almost all patients were on ICT, with roughly half of
2/3), and ketoacidosis (3/3 and 5/3).
the patients injecting insulin four to five times per day, and nearly
one-third of patients injecting insulin six times per day in both
groups. The most frequently reported indications for transition to 4 | DI SCU SSION
CSII were unpredictable swings in blood glucose concentration, the
wish for flexibility in daily routine and patient preference with a In the pumpkin study, the self-reported DHRQOL in school-children
nearly equal proportion in both the CSII and the MDI group. HbA1c- aged 8 to 11 years improved 6 months after transition to CSII com-
values at baseline were generally satisfying, and 0.5% lower in the pared to those who stayed on MDI. In adolescents aged 12 to
CSII group. 16 years, there was no significant difference between the treatment
MUELLER-GODEFFROY ET AL. 5

Assessed for eligibility (n = 367)

Not included (n = 156)


Not meeting inclusion criteria (n = 39)
Declined to participate (n = 74)
Other reasons (n = 26)
Unknown reasons (n = 17)

Randomized (n = 211)

Allocated to CSII (n = 106) Allocated to MDII (n = 105)


Received intervention (n = 104) Received intervention (n = 100)
Age group 6 - 7 years: 10 Age group 6 - 7 years: 8
Age group 8 - 11 years: 43 Age group 8 - 11 years: 36
Age group 12 - 16 years: 51 Age group 12 - 16 years: 56

Did not receive intervention (n = 2) Did not receive intervention (n = 5)


Age group 5 - 7 years: Age group 12 - 16 years:
2 Rejection of CSII 4 Non-Compliance
1 Logistical problems in center

6 month follow-up 6 month follow-up


Returned for follow up (n = 90) Returned for follow up (n = 84)
Age group 6 - 7 years: 9 Age group 6 - 7 years: 7
Age group 8 - 11 years: 37 Age group 8 - 11 years: 28
Age group 12 - 16 years: 44 Age group 12 - 16 years: 49

Did not return for follow-up (n = 14) Did not return for follow-up (n = 16)
Age group 6 - 7 years: Age group 6 - 7 years:
1 CRF overwhelming 1 Non-Compliance
Age group 8 - 11 years: Age group 8 - 11 years:
3 Logistical problems in center 3 Logistical problems in center
2 Rejection of CSII 2 Rejection of CSII
1 Unknown reasons 2 Unknown reasons
Age group 12 - 16 years: 1 Medical reasons
4 Logistical problems in center Age group 12 - 16 years:
3 Non-compliance 4 Non-Compliance
3 Logistical problems in center

Analyzed FAS (n = 90) Analyzed FAS (n = 89)


Age group 6 - 7 years: 9 Age group 6 - 7 years: 7
Age group 8 - 11 years: 35 Age group 8 - 11 years: 31
Age group 12 - 16 years: 46 Age group 12 - 16 years: 51

Secondary endpoint available Secondary endpoint available


Age group 6 - 7 years: 10 Age group 6 - 7 years: 8
Age group 8 - 11 years: 40 Age group 8 - 11 years: 33
Age group 12 - 16 years: 48 Age group 12 - 16 years: 60

FIGURE 1 Participant flow

groups. Overall burden on the main caregiver decreased significantly (without age-stratification) showed significantly improved DHRQOL
in the CSII group compared to the MDI group. in the CSII group compared to the MDI group in children and adoles-
The age-stratified approach to measuring children's and adoles- cents aged 8 to 16 years. This result is in accordance with those of
cents' DHRQOL of the pumpkin trial makes it difficult to compare the some recently published large cross-sectional or registry-based stud-
results with other studies. The analysis of the overall pumpkin sample ies based on a wide age-range.28,29 Two studies employing an age-
6 MUELLER-GODEFFROY ET AL.

TABLE 1 Baseline characteristics of the study participants

n CSII n MDI
Age (years) 98 11.3  2.7 101 11.9  2.8
Male sex—no. (row %) 98 56 (49.6) 101 57 (50.4)
HbA1c (mmol/Mol) (%) 95 56.3  9.8 (7.3  0.9) 100 61.7  14.2 (7.8  1.3)
Diabetes duration (years) 91 3.3  2.9 92 3.6  3.0
Injections per day—no. (row %) 96 — 100 —
3 — 2 (100.0) — 0
4-5 — 49 (48.9) — 51 (51.1)
6 — 34 (48.6) — 36 (51.4)
7-8 — 11 (42.1) — 13 (57.9)
Indication for transition to CSIIa
Improvement of HbA1c — 24 (42.1 — 33 (57.9)
Dawn/dusk-phenomenon — 46 (47.4) — 51 (52.6)
Unpredictable swings in blood glucose concentrations — 61 (54.0) — 52 (46.0)
Recurrent hypoglycemia — 35 (54.7) — 29 (45.3)
Flexibility in mealtime/irregular daily routine — 61 (48.4) — 65 (51.6)
Patient preference — 52 (45.6) — 62 (54.4)
Socioeconomic status - no. (row %) 85 — 83 —
High — 35 (56.5) — 27 (43.5)
Medium — 37 (48.7) — 39 (51.3)
Low — 13 (43.3) — 17 (56.7)

Plus-minus values are means  SD.


a
More than one response allowed.

TABLE 2 Primary outcomes at baseline and 6 months follow-up

CSII MDI CSII vs MDI


n Mean SD a n Mean SD a Med diffb (95% CI) Pc Pd
Patient self-report
Diabetes specific quality of life Children 8-11 y
KINDL-DM (range 0-100)
Baseline 36 68.1 14.9 29 61.8 15.2 — — —
Follow-up 35 74.5 12.0 31 64.3 14.9 9.5 (3.6 to 16.7) 0.004 0.001
Adolescents 12-16 y
Baseline 46 70.6 11.9 57 67.8 16.9 — — —
Follow-up 46 74.2 13.0 53 70.9 16.0 2.7 (−3.2 to 9.5) 0.353 0.606
Main caregiver report Total sample
Overall diabetes Baseline 92 3.4 1.0 94 3.2 1.1 — — —
burden (range 1-5)
Follow-up 90 2.7 1.1 89 3.0 1.1 0 (−1 to 0) 0.029 0.005

Abbreviations: CI, confidence interval; CSII, continuous subcutaneous insulin infusion; KINDL-DM, diabetes-specific module of the KINDL-R; KINDL-R,
modular HRQOL questionnaire for children-revised version; MDI, multiple daily injections.
a
Standard deviations.
b
Difference of the median.
c
Exact two-sided U-test.
d
Asymptotic test after adjustment for baseline and age and with center as a random factor.

stratified approach10,30 showed—in contrast with our results—better RCT in children <14 years33 found no difference in DHRQOL
DHRQOL associated with CSII compared to MDI in adolescents. between CSII and MDI. An RCT in adolescents aged 14 to 18 years
These studies examined pump users as a group. CSII users may repre- reported better DHRQOL in those who used CSII.34 The differences
sent a cluster with specific characteristics associated with better to our results might be owing to different age-ranges, the use of
HRQOL independent of pump use. To analyze whether the individual different instruments, and also the small sample size of the older
patient benefits, studies on patient data before and after transition to RCTs (n = 16-38).
the insulin pump compared to a—preferably randomized—control Adolescent-reported diabetes treatment satisfaction improved in
group are indicated. both treatment groups, with the difference in favor of CSII not meet-
RCTs on the impact of CSII on DHRQOL in pediatric diabetes ing the predefined significance level for secondary outcomes. The
are rare and were published 10 years ago or more. Contrary to our improvement in the waiting group may be owing to the optimization
findings, two RCTs including a wide age-range,31,32 as well as an of MDI treatment at the beginning of the study, and also may reflect
MUELLER-GODEFFROY ET AL. 7

TABLE 3 Secondary outcomes at baseline and 6 months follow-up

CSII MDI CSII vs MDI


Age group n Mean SD a
n Mean SD a
Med diffb (95% CI) Pc Pd
Patient self-report
Health-related quality of life Children 8-11 y
KINDL-TS
Baseline 36 73.3 9.9 29 73.4 7.6 — — —
(range 0-100)
Follow-up 35 72.3 10.3 31 71.7 8.7 0.2 (−4.2 to 5.2) 0.881 0.590
Adolescents 12-16 y
Baseline 46 72.7 8.0 56 69.7 9.3 — — —
Follow-up 46 70.6 12.0 52 71.2 8.2 0 (−3.1 to 4.2) 0.863 0.440
Family conflict Adolescents 12-16 y
DFCS
Baseline 46 25.8 4.1 56 27.3 6.2 — — —
(range 19-57)
Follow-up 46 25.7 4.9 52 26.6 6.6 0 (−1.9 to 1) 0.893 0.859
Treatment satisfaction Adolescents 12–16 y
DTSQ-TS + CTeen
Baseline 46 39.4 9.1 57 36.9 8.7 — — —
(range 0-54
Follow-up 46 44.4 7.7 53 40.5 9.7 4 (1 to 6) 0.015 0.042
Main caregiver report
Total sample
HRQOL, WHO-5 (range 0–100) Baseline 92 51.9 24.5 91 50.3 22.2 — — —
Follow-up 87 59.4 19.2 89 53.4 20.3 4 (0 to 12) 0.056 0.103
Parenting stress, PIP TS-F Baseline 91 96.9 22.7 92 95.7 23.2
(range 42-210)
Follow-up 87 83.3 21.1 87 91.6 24.7 −8 (−16 to −1) 0.020 0.001
PIP TS-D (range 42-210) Baseline 88 90.1 25.8 88 89.7 24.1 — — —
Follow-up 82 77.4 22.8 85 85.5 26.7 −7 (−15 to 0) 0.047 0.004
Fear of hypoglycemia, Baseline 91 30.0 5.3 93 30.2 6.1 — — —
HFS behavior (range 10-50)
Follow-up 88 26.0 5.7 89 28.5 6.1 −2 (−4 to 1) 0.006 0.001
HFS worry (range 15-75) Baseline 92 40.8 10.5 94 41.3 11.0 — — —
Follow-up 90 35.2 10.6 89 40.4 12.2 −5 (−8 to 1) 0.004 <0.001
Family conflict, DFCS Adolescents 12–16 y
mea (range 19–57)
Baseline 46 28.3 7.2 57 30.0 6.6 — — —
Follow-up 46 26.8 6.7 50 29.6 7.8 −3 (−5 to 0) 0.046 0.063
Treatment satisfaction, Total sample
DTSQ- TS + C P (range 0-66)
Baseline 92 44.0 10.5 93 41.6 10.6 — — —
Follow-up 90 52.9 8.8 88 46.3 9.6 7 (4 to 9) <0.001 <0.001
HbA1c, mmol/mol (%) Young children 6–7 y
Baseline 9 53.0 (7.0) 8.7 (0.8) 8 55.2 (7.2) 8.7 (0.8) — — —
Follow-up 10 53.0 (7.0) 5.5 (0.5) 8 54.1 (7.1) 7.7 (0.7) −0.2 (−0.9 to 0.4) 0.530 0.602
Children 8–11 y
Baseline 40 55.2 (7.2) 8.7 (0.8) 33 58.5 (7.5) 12.0 (1.1) — — —
Follow-up 40 54.1 (7.1) 10.9 (1.0) 33 59.6 (7.6) 12.0 (1.1) −0.4 (−0.8 to 0.1) 0.085 0.204
Adolescents 12–16 y
Baseline 46 57.4 (7.4) 12.0 (1.1) 59 61.7 (7.8) 16.4(1.5) — — —
Follow-up 46 56.3 (7.3) 10.9 (1.0) 55 61.7 (7.8) 14.2 (1.3) −0.5 (−0.9 to −0.1) 0.025 0.077

Abbreviations: CI, confidence interval; CSII, continuous subcutaneous insulin infusion; DFCS, diabetes family conflict scale; HFS, hypoglycemia fear survey;
HRQOL, health-related quality of life; KINDL-DM, diabetes-specific module of the KINDL-R; KINDL-R, modular HRQOL questionnaire for children-revised
version; MDI, multiple daily injections; PIP, pediatric inventory for parents; WHO-5, World Health Organization-five.
a
Standard deviations.
b
Difference of the median.
c
Exact two-sided U-test.
d
Asymptotic test after adjustment for baseline and with center as a random factor.

the fact that patients may experience some benefits, for example, in Adolescents with T1DM form a challenging patient group. They
terms of hope or positive expectations, even before they receive the are confronted with biological and psychosocial changes which may
intended treatment change. This may have reduced the effect threaten their metabolic control, mental health, and HRQOL. They
between the treatment groups. All in all, neither DHRQOL, nor treat- have to take progressively more self-responsibility for their treatment,
ment satisfaction improved as much as we expected in adolescents. and both taking premature responsibility and the over-involvement of
8 MUELLER-GODEFFROY ET AL.

the parents may impact negatively on the adolescents' metabolic con- group, which may have reduced the differences observed between
35
trol and HRQOL. Self-management activities and the need to wear the treatment groups.
insulin administration devices may induce a feeling of being different The DHRQOL of the adolescents in the pumpkin trial did not
from peers.36,37 Also, adolescents may understand for the first time improve, while their caregivers reported a decline of their own diabe-
the implications of having a lifelong disease. These issues inherent to tes burden. This study did not examine whether the diabetes manage-
the adolescents' developmental stage may result in a lower increase in ment tasks between child and caregivers changed after the transition
DHRQOL than with school-aged children. In order to improve adoles- to CSII, and how satisfied caregivers and adolescents were with those
cents' DHRQOL, additional efforts beyond the transition to CSII are eventual changes. Future studies are necessary to understand the pro-
needed. cesses leading to the findings observed.
Childhood diabetes is a profound stressor for all family members.
As other studies, we found a decline of caregiver stress,12,23 and fear
about hypoglycemia38,39 associated with CSII. These results of our 5 | CONC LU SIONS
trial, representing different aspects of how the care for a child with
diabetes may impact life and well-being of the caregivers, clearly CSII has substantial psychosocial benefits, as it increases pre-
favored CSII over MDI. The main caregivers in the CSII group were adolescent children's diabetes-specific quality of life and decreases
highly satisfied with their child's diabetes regimen, significantly more caregiver burden. Patient-reported outcome measures help to under-
so than those whose child stayed on MDI. High satisfaction with CSII stand the patients' and parents' experience.
is also reported by other studies,40 among these RCTs.21,33
The initial HbA1c values of the study sample as a whole were sat-
ACKNOWLEDGEMENTS
isfactory. Changes in HbA1c were small, and there was no difference
between the treatment groups. This large-scale RCT showed no signif- The authors thank the investigators and diabetes teams for their
icant differences between treatment groups in acute complications. cooperation: Angela Galler, Pediatric Diabetes Center of the Univer-
Acute complications are infrequent events, and they may not have sity Clinics Charité Berlin; Burkhardt Brosig, Miriam Kramer, Pediatric
occurred during the comparatively short study period. Endocrine Clinic of the Justus Liebig University Gießen; Paul-Martin
Despite the thorough randomization, the initial self-reported Holterhus, Department of Pediatric and Adolescent Medicine of the
DHRQOL and some secondary outcomes differed between the treat- University Clinic of Schleswig-Holstein, Campus Kiel; Simone von
ment groups to a clinically relevant degree. An analysis of covariance Sengbusch, Department of Pediatric and Adolescent Medicine of the
was conducted to adjust for baseline differences. Furthermore, as University Clinic of Schleswig-Holstein, Campus Luebeck; Ulrike
these initial dissimilarities favored the CSII group and nonetheless the Jacobi, Pediatric University Clinic Rostock; Andreas Neu, Franziska
patients reported higher improvements compared to the MDI group, Liebrich, Pediatric University Clinic Tübingen; Markus Freff, Children's
these imperfections of randomization did not weaken our findings. In Hospitals PRIMA Darmstadt; Ulrike Menzel, Children's Hospital
general, the RCT pumpkin confirmed the benefits of CSII described in Hamburg-Altona; Kirsten Mönkemöller, Children's Hospital Köln
the pilot study, albeit with smaller effects, and not in all age groups. Amsterdamer Strasse; Susanne Büsing, Christian Children's Hospital
The effects in RCTs tend to be smaller than those in observational Osnabrück; Eva Hahn, Protestant Hospital Oberhausen; Martin
pre-post trials. Patients may have been not included, either because
Holder, Olgahospital Stuttgart; Marianne Becker, HELIOS Clinic Wies-
they themselves were not willing to wait or because they were recom-
baden; Beate Karges, Bethlehem Health Center Stolberg; Germany.
mended for immediate transfer to CSII for medical reasons. Precisely
The authors also thank their diabetes teams, and Reinhard Holl, Bele
these patients may be expected to benefit strongly from CSII, and
Jacobs, and Ruediger Szczepanski (Data Safety and Monitoring Com-
thus, the RCT may underestimate the benefit of CSII in routine care.11
mittee) for their support, and all patients and families who contributed
to the trial. This trial was mainly funded by the German Research
4.1 | Strengths and limitations of the study Foundation (WA 2929/2-1), with additional financial support by
We regard our study as the largest RCT to date which assesses the Roche Diagnostics GmbH, Diabetes Care, Germany.
impact of CSII on HRQOL and diabetes burden in the pediatric age
group. The age-stratified analysis of child self-reported outcomes, the
AUTHOR CONT RIB UTI ON S
use of CSII-adapted and previously tested instruments and the partici-
pation of 18 certified diabetes centers add further strength to this The agip, C.L.S., B.H., C.B., D.H., and E.L contributed research data.
trial. E.M.G., V.W., N.H., and R.V. wrote the manuscript. R.V., E.M.G., and
However, because of the already common use of CSII in Ger- A.Z. planned the analysis. N.H. and R.V. conducted the analysis.
many, the recruitment time was much longer than expected. And in E.M.G., V.W., N.H., R.V., and A.Z. reviewed/edited the manuscript. All
the age group <8 years, the number of patients was too small to reli- contributed significantly to discussion.
ably detect differences. The authors acknowledge the programming assistance and data
The waiting-list design of the pumpkin trial ensured that only clin- monitoring of Frank Sandig, IMBS, University of Luebeck, and Sabine
ically comparable patients—those with an indication for CSII—were Brehm, Department of Pediatric and Adolescent Medicine, University
included. However, we found some positive effects in the waiting of Luebeck, and the monitoring of ZKS-Luebeck guarantee.
MUELLER-GODEFFROY ET AL. 9

CONF LICT S OF INTE R ES T Bundesgesundheitsblatt Gesundheitsforschung Gesundheitsschutz. 2007;


50(5–6):578-589.
V.W. received a grant by Roche Diagnostics GmbH, Diabetes Care, 17. Ravens-Sieberer U, Bullinger M. KINDL-R Manual 2000 Hamburg, Ger-
Germany for presenting data at a symposium at the ISPAD 2017, Inns- many. http://kindl.org/english/manual/.
bruck, Austria, B.H. received a grant by Ipsen for the ESPE meeting 18. Anderson BJ, Brackett J, Ho J, Laffel LM. An office-based intervention
to maintain parent-adolescent teamwork in diabetes management.
2017. All other authors declare no conflict of interest.
Impact on parent involvement, family conflict, and subsequent glyce-
mic control. Diabetes Care. 1999;22(5):713-721.
ORCID 19. Hood KK, Butler DA, Anderson BJ, Laffel LMB. Updated and revised
diabetes family conflict scale. Diabetes Care. 2007;30(7):1764-1769.
Esther Mueller-Godeffroy http://orcid.org/0000-0003-3237-9556 20. Bradley C, Lewis KS. Measures of psychological wellbeing and treat-
ment satisfaction developed from the responses of people with
tablet-treated diabetes. Diabet Med. 1990;7:445-451.
RE FE R ENC E S 21. Skogsberg L, Fors H, Hanas R, Chaplin JE, Lindman E, Skogsberg J.
1. Danne T, Bangstad H-J, Deeb L, et al. Insulin treatment in children and Improved treatment satisfaction but no difference in metabolic control
adolescents with diabetes. Pediatr Diabetes. 2014;15:115-134. when using continuous subcutaneous insulin infusion vs. multiple daily
2. Sherr JL, Hermann JM, Campbell F, et al. Use of insulin pump therapy injections in children at onset of type 1 diabetes mellitus. Pediatr Dia-
in children and adolescents with type 1 diabetes and its impact on betes. 2008;9:472-479.
metabolic control. Diabetologia. 2016;59(1):87-91. 22. WHO-5 Questionnaires. http//www.psykiatri-regionh.dk
3. Misso ML, Egberts KJ, Page M, O’Connor D, Shaw J. Continuous sub- 23. Streisand R, Swift E, Wickmark T, Chen R, Holmes CS. Pediatric par-
cutaneous insulin infusion (CSII) versus multiple insulin injections for enting stress among parents of children with type 1 diabetes: the role
type 1 diabetes mellitus. Cochrane Database Syst Rev. 2010;(1): of self-efficacy, responsibility, and fear. J Pediatr Psychol. 2005;30:
CD005103. 513-521.
4. Karges B, Schwandt A, Heidtmann B, et al. Association of Insulin Pump 24. Clarke WL, Gonder-Frederick LA, Snyder AL, Cox DJ. Maternal fear of
Therapy vs insulin injection therapy with severe hypoglycemia, ketoa- hypoglycemia in their children with insulin dependent diabetes melli-
tus. J Pediatr Endocrinol Metab. 1998;11(1):189-194.
cidosis, and glycemic control among children, adolescents, and young
25. Haugstvedt A, Wentzel-Larsen T, Aarflot M, Rokne B, Graue M.
adults with type 1 diabetes. JAMA. 2017;318(14):1358-1366.
Assessing fear of hypoglycemia in a population-based study among
5. Yeh HC, Brown TT, Maruthur N, et al. Comparative effectiveness and
parents of children with type 1 diabetes - psychometric properties of
safety of methods of insulin delivery and glucose monitoring for diabe-
the hypoglycemia fear survey - parent version. BMC Endocr Disord.
tes mellitus: a systematic review and meta-analysis. Ann Intern Med.
2015;15(1):2.
2012;157:336-347.
26. The Diabetes Control and Complications Trial Research Group,
6. Pate T, Klemencic S, Battelino T, et al. Fear of hypoglycemia, anxiety,
Nathan DM, Genuth S, et al. The effect of intensive treatment of dia-
and subjective well-being in parents of children and adolescents with
betes on the development and progression of long-term complications
type 1 diabetes. J Health Psychol. 2016;(1):135910531665093.
in insulin-dependent diabetes mellitus. N Engl J Med. 1993;329:
https://doi.org/10.1177/1359105316650931.
977-986.
7. Reaney M, Black P, Gwaltney C. A systematic method for selecting
27. Bretz F, Maurer W, Brannath W, Posch M. A graphical approach to
patient-reported outcome measures in diabetes research. Diabetes
sequentially rejective multiple test procedures. Stat Med. 2009;28(4):
Spectr. 2014;27:229-232.
586-604.
8. Hirose M, Beverly EA, Weinger K. Quality of life and technology:
28. Birkebaek NH, Kristensen LJ, Mose AH, Thastum M, Danish Society
impact on children and families with diabetes. Curr Diab Rep. 2012;
for Diabetes in Childhood and Adolescence. Quality of life in Danish
12(6):711-720.
children and adolescents with type 1 diabetes treated with continuous
9. Graue M, Wentzel-Larsen T, Hanestad BR, Batsvik B, Sovik O. Mea-
subcutaneous insulin infusion or multiple daily injections. Diabetes Res
suring self-reported, health-related, quality of life in adolescents with
Clin Pract. 2014;106(3):474-480.
type 1 diabetes using both generic and disease-specific instruments. 29. Lawrence JM, Yi-Frazier JP, Black MH, et al. Demographic and clinical
Acta Paediatr. 2003;92:1190-1196. correlates of diabetes-related quality of life among youth with type
10. Stahl-Pehe A, Straßburger K, Castillo K, et al. Quality of life in inten- 1 diabetes. J Pediatr. 2012;161(2):201-207.
sively treated youths with early-onset type 1 diabetes: a 30. Cherubini V, Gesuita R, Bonfanti R, et al. Health-related quality of life
population-based survey. Pediatr Diabetes. 2014;15:436-443. and treatment preferences in adolescents with type 1 diabetes. The
11. Blair JC, Peak M, Gregory JW. What is the best way to deliver subcu- VIPKIDS study. Acta Diabetol. 2014;51(1):43-51.
taneous insulin to infants, children, and young people with type 1 dia- 31. Nuboer R, Borsboom GJJM, Zoethout JA, Koot HM, Bruining J. Effects
betes mellitus? BMJ. 2011;343:d5221. of insulin pump vs. injection treatment on quality of life and impact of
12. Barnard KD, Lloyd CE, Skinner TC. Systematic literature review: qual- disease in children with type 1 diabetes mellitus in a randomized, pro-
ity of life associated with insulin pump use in type 1 diabetes. Diabet spective comparison. Pediatr Diabetes. 2008;9(4.1:291-296.
Med. 2007;24:607-617. 32. Doyle EA, Weinzimer SA, Steffen AT, Ahern JAH, Vincent M,
13. Nieuwesteeg A, Pouwer F, van der Kamp R, van Bakel H, Aanstoot HJ, Tamborlane WV. A randomized, prospective trial comparing the efficacy
Hartman E. Quality of life of children with type 1 diabetes: a system- of continuous subcutaneous insulin infusion with multiple daily injec-
atic review. Curr Diabetes Rev. 2012;8(6):434-443. tions using insulin glargine. Diabetes Care. 2004;27(7):1554-1558.
14. Müller-Godeffroy E, Treichel S, Wagner VM. Investigation of quality 33. Weintrob N, Benzaquen H, Galatzer A, et al. Comparison of continu-
of life and family burden issues in insulin pump therapy in children ous subcutaneous insulin infusion and multiple daily injection regi-
with type 1 diabetes mellitus – a large scale multicenter pilot study. mens in children with type 1 diabetes: a Randomizzed open crossover
Diabet Med. 2009;26:493-501. trial. Pediatrics. 2003;112(3):559-564.
15. Neu A, Beyer P, Bürger-Büsing J, et al. Diagnosis, therapy and control 34. Cohen D, Weintrob N, Benzaquen H, Galatzer A, Fayman G, Phillip M.
of diabetes mellitus in children and adolescents. Exp Clin Endocrinol Continuous subcutaneous insulin infusion versus multiple daily injec-
Diabetes. 2014;122:425-434. tions in adolescents with type I diabetes mellitus: a randomized open
16. Lange M, Kamtsiuris P, Lange C, Schaffrath Rosario A, Stolzenberg H, crossover trial. J Pediatr Endocrinol Metab. 2003;16(7):1047-1050.
Lampert T. Sociodemographic characteristics in the German health 35. Tsiouli E, Alexopoulos EC, Stefanaki C, Darviri C, Chrousos GP. Effects
interview and examination survey for children and adolescents of diabetes-related family stress on glycemic control in young patients
(KiGGS)—operationalisation and public health significance, taking as with type 1 diabetes: systematic review. Can Fam Physician. 2013;
an example the assessment of general state of health. 59(2):143-149.
10 MUELLER-GODEFFROY ET AL.

36. Almeida AC, Leandro E, da Graca Pereira M. The role of family func- SAGE Open Medicine. 2017;5:2050312117694938, 20503121176
tioning on metabolic control and quality of life in adolescents with 9493.
type 1 diabetes mellitus. In: Nunes KP, ed. (eds) Major Topics in Type
1 Diabetes. InTechOpen, 2015, London. SUPPOR TI NG I NFORMATION
37. Borus JS, Laffel L. Adherence challenges in the management of type
1 diabetes in adolescents: prevention and intervention. Curr Opin Additional supporting information may be found online in the Sup-
Pediatr. 2010;22(4):405-411. porting Information section at the end of the article.
38. Johnson SR, Cooper MN, Davis EA, Jones TW. Hypoglycaemia, fear of
hypoglycaemia and quality of life in children with Type 1 diabetes and
their parents. Diabet Med. 2013;30:1126-1131. How to cite this article: Mueller-Godeffroy E, Vonthein R,
39. Nixon R, Pickup JC. Fear of hypoglycemia in type 1 diabetes managed
Ludwig-Seibold C, et al. Psychosocial benefits of insulin pump
by continuous subcutaneous insulin infusion: is it associated with poor
glycemic control? Diabetes Technol Ther. 2011;13:93-98. therapy in children with diabetes type 1 and their families: The
40. Hussain T, Akle M, Nagelkerke N, Deeb A. Comparative study on pumpkin multicenter randomized controlled trial. Pediatr Dia-
treatment satisfaction and health perception in children and ado-
betes. 2018;1–10. https://doi.org/10.1111/pedi.12777
lescents with type 1 diabetes mellitus on multiple daily injection
of insulin, insulin pump and sensor-augmented pump therapy.

Anda mungkin juga menyukai